Comparison of Escherichia coli surface attachmentmethods for single-cell, in vivo microscopyYao Kuan Wang1, +, Ekaterina Krasnopeeva2, +, Ssu-Yuan Lin1, Fan Bai3, Teuta Pilizota2, *,and Chien-Jung Lo1, *

1Department of Physics and Graduate Institute of Biophysics, National Central University, Jhongli, Taiwan 32001,ROC2Centre for Synthetic and Systems Biology, Institute of Cell Biology, School of Biological Sciences, University ofEdinburgh, Alexander Crum Brown Road, EH9 3FF, Edinburgh, UK3Biodynamic Optical Imaging Center (BIOPIC), School of Life Sciences, Peking University, Beijing 100871, China*corresponding authors teuta.pilizota@ed.ac.uk, cjlo@phy.ncu.edu.tw+these authors contributed equally to this work

ABSTRACT

For in vivo, single-cell imaging bacterial cells are commonly immobilised via physical confinement or surface attachment.Different surface attachment methods have been used both for atomic force and optical microscopy (including super resolution),and some have been reported to affect bacterial physiology. However, a systematic comparison of the effects these attachmentmethods have on the bacterial physiology is lacking. Here we present such a comparison for bacterium Escherichia coli,and assess the growth rate, size and intracellular pH of cells growing attached to different, commonly used, surfaces. Wedemonstrate that E. coli grow at the same rate, length and internal pH on all the tested surfaces when in the same growthmedium. The result suggests that tested attachment methods can be used interchangeably when studying E. coli physiology.

IntroductionMicroscopy has been a powerful tool for studying biological processes on the cellular level ever since the first discovery ofmicroorganisms by Antonie van Leeuwenhoek back in 17th century1. Recently employed in vivo single-cell imaging allowedscientists to study population diversity2, physiology3, sub-cellular features4, and protein dynamics5 in real-time. Single cellimaging of bacteria is particularly dependent on immobilisation, as majority of bacteria are small in size and capable ofswimming. Immobilisation methods vary depending on the application, but typically fall into one of the two categories: use ofphysical confinement or attachment to the surface via specific molecules. The former group includes microfluidic platformscapable of mechanical trapping6, 7, where some popular examples include the ”mother machine”8, CellASIC9 or MACS10

devices, and porous membranes such as agarose gel pads2, 11–13. Physical confinement methods, while higher in throughout,can have drawbacks. For example, agarose gel pads do not allow fast medium exchange, and mechanically constrained bacteriacannot be used for studies of bacterial motility or energetics via detection of bacteria flagellar motor rotation14–16. Chemicalattachment methods rely on the interaction of various adhesive molecules, deposited on the cover glass surface, with thecell itself. Adhesion can be a result of electrostatic (polyethylenimine (PEI)17, 18, poly-L-lysine (PLL)14, 16, 18) or covalentinteractions (3-aminopropyltriethoxysilane (APTES)18), or a combination, such as with polyphenolic proteins (Cell-TakT M)18.

Time scales on which researchers perform single-cell experiments vary. For example, scanning methods, like atomic forcemicroscopy (AFM) or confocal laser scanning microscopy (CLSM), require enough time to probe each point of the sample,and stochastic approaches of super resolution microscopy (e.g. PALM and STORM) use low activation rate of fluorophoresto achieve a single fluorophore localisation. Thus, required acquisition time scales vary from milliseconds to minutes19, 20.Experiments aimed at the observation of cell growth or slow cellular responses can run from minutes to hours8, 11, 15, 21, 22.

Regardless of the time scale, physiology of the studied bacteria should not be affected by the surface attachment. Forexample, single particle tracking is often performed on surface immobilised cells5, 23, 24, and cellular physiology can influenceparticle diffusion in the cytoplasm, e.g. metabolic ”stirring” of the cytoplasm enhances diffusion in size dependant manner25,and in yeasts an intracellular pH has been shown to affect cytoplasm fluidity26. Furthermore, concerns have been raised thatcharged molecules, like PLL that is commonly used for a surface attachment16, 22, 23, 27, 28, can perturb membrane potentialcausing partial or complete membrane depolarisation29, 30. Additionally, PLL in high concentrations exhibits antimicrobialproperties31. Despite of these concerns, PLL has been widely used in super-resolution and single molecules tracking applicationsas it is cheap and easy to use32, 33. It is, therefore, important to characterise physiological parameters on different surfaces, and

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

on the time scales relevant for live cell imaging.In this report we compare a range of immobilisation techniques, including PLL, PEI, Cell-TakT M and agarose gel pad, using

Escherichia coli as a model organism. We measure several physiological traits during growth on the specific surface, includinggrowth rate, size and intracellular pH, and find that tested immobilisation methods do not differ; growth rate and cell size aresurface-attachment independent.

Results

Immobilisation assaysWe test four substrates commonly used for bacteria immobilisation: poly-L-lysine (PLL)14, 30, polyethylenimine (PEI)17, 34,Cell-TakT M18, 3-aminopropyltriethoxysilane (APTES)18 and agarose gel20.

PLL and PEI are cationic polymers, which can electrostatically interact with negative charges on the outer surface ofthe cell35. Several PLL coating protocols have been reported, which we here refer to as ”in-chamber”14, ”rinsed”30, and”air-dried”30 methods. ”In-chamber” PLL coating is the most standard for bacterial flagellar motor experiments commonlyknown as bead assay14, 36–38. In this protocol PLL solution is flushed into an uncoated glass flow-chamber for 10-15 s followedby thorough washing with the excessive volume of growth medium (∼ 25 times the flow-chamber volume). In the ”rinsed”method lower PLL concentration and longer incubation time (min) are used to cover the entire surface of the coverslip byimmersing it in the PLL solution30 amd subsequently washing. ”Air-dried” method is similar to the ”rinsed”, with an addition ofdrying the PLL solution on the coated surface for over an hour. For our detailed coating protocols see Materials and Methods.

Cell-TakT M is a commercially available adhesive extracted from marine mussel, Mytilus edulis. It’s a component of byssus,a bundle of filaments mussels secrete to anchor themselves to solid surfaces39, 40, where characterising and mimicking theadhesive chemistry of mussel byssus is an active area of research41. What we know from these efforts thus far, is that it involvesbidentate and covalent interactions, protein coacervation, intrinsic protein-protein binding as well as metal chelation41. We usethe manufacturer coating protocol as described in Materials and Methods.

APTES is a common choice for salinisation of microfluidic channels42, and has also been employed as an attachment agentfor the AFM imaging18. We coat coverslips by incubating them in a 2% solution of APTES for 2 h followed by extensivewashing with water and acetone as described in Materials and Methods.

As a control we grow bacteria on the agarose gel pad with no chemical adhesives (see Materials and Methods for details).

A

PLLin-chamber

PLLrinsed

PLLair-dried

PEI Cell-Tak Gel pad

0.000

0.005

0.010

0.015

0.020

0.025

0.030

Gro

wth

rate

(1/m

in)

B

Figure 1. Bacterial growth rate stays constant when cells are attached to the surface with different attachment methods. (A)The single cell length is measured using phase contrast imaging over a complete cell cycle (from the first to the second division)and fitted with a single exponential function to obtain a single cell growth rate. Panel on the right show examples of the phasecontrast images at the indicated time points. Scale bar is 2 µm. (B) Comparison of average growth rates on different surfacesshows no significant difference between all tested immobilisation methods in MM9 medium. The growth rates were measuredat 23C. Number of cells analysed for each attachment method is: N=44, 104, 28, 81, 48, and 43 for ”PLL in-chamber”, ”PLLrinsed”, ”PLL air-dried”, PEI, Cell-Tak and gel pad respectively.

2/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

Growth rate and morphology of bacteria do not dependent on the surface attachment methodTo estimate the effect of different immobilisation assays on bacterial physiology we first examine the growth rate and thecells’ morphology as the cells grow on each specific surface8. To obtain the growth rate and information on morphologywe monitor bacteria between first and second divisions using optical microscopy (see Materials and Methods). Bacterialgrowth assays are performed at 23C in the flow chambers, where fresh oxygenated medium is constantly supplied to the cells.Phase-contrast images of the bacteria are taken every 5 min and cells’ length is extracted as described in Materials and Methodsand Figure 1A. Figure 1B shows single cell growth rates, which we define as a rate of relative elongation of an exponentiallygrowing bacterium8, 43, of the cells grown on different surfaces. The growth rate observed is independent on the surfaceattachment, and, when we average the values obtain for the 6 different surfaces in Figure 1B, equals (0.0071 ± 0.0044 min−1) .As expected, the growth rate is medium dependant and becomes twice as high when we move from the MM9 medium (seeMaterials and Methods) to the richer LB medium (0.0153 ± 0.0049 min−1), SI Figure 2A.

In addition to the growth rates, in Figure 2A we analyse the lengths of individual cells at the beginning and at the endof the growth cycle. Similarly to the growth rates, the initial (L0) and final (L f ) length, and cell width (W ) are maintainedconstant for different immobilisation methods: L0 = 2.61±0.31µm in MM9 (Figure 2A) and 3.31 ± 0.54 µm in LB medium(SI Figure 2B), L f = 4.35±0.49 µm in MM9 (Figure 2A inset) and 5.72±0.95 µm in LB, and W = 0.94±0.04 µm in MM9(Figure 2B), and 1.03±0.07 µm in LB (SI Figure 2D). The length ratio (defined as the L f /L0) stays 1.69±0.21, irrespectiveof the immobilisation method or the growth media (SI Figure 1).

Though independent of the immobilisation protocol, the length of the bacteria is sensitive to the surface attachment. Whenattached to a surface, average cell length grows smaller with time reaching a steady level already after the first division (SIFigure 3). The mean of the length distribution of surface-grown bacteria (after first division) is approximately 30% smaller thanthat of the planktonic cell population.

PLLin-chamber

PLLrinsed

PLLair-dried

PEI Cell-Tak Gel pad1

2

3

4

5

6

7

8

9

Init

ial le

ngth

(m

)

PLLin-chamber

PLLrinsed

PLLair-dried

PEI Cell-Tak Gel pad0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

Wid

th (

m)

0

2

4

6

8

Final le

ngth

(m

)A B

Figure 2. Morphology of bacteria is not influenced by the attachment method. (A) Initial cell length on different surfaces.Cells are grown in a bulk culture and immobilised on the surface. The length is measured after the first division followingimmobilisation. The inset shows the final cell length before the second division. (B) Width distribution of cells attached to thesurface with various immobilisation methods remains constant.

Intracellular pH during growth on the surface does not depend on the method of attachmentNeutrophilic bacteria maintain their cytoplasmic pH withing a narrow range (termed pH homeostasis). For example, E. colican survive in a range of external pHs, starting as low as pH ∼ 2 in the human stomach and up to pH ∼ 9 at the pancreaticduct, while maintaining internal pH in a relatively narrow range of 7-844–49. Cytoplasmic pH plays an important role in cellularenergetics as the difference between cytoplasmic and extracellular pH contributes to the electrochemical gradient of protons(so called proton motive force50), as well as influences protein stability and an enzymatic activity in the cell51. However,cytoplasmic pH can change when cells are subjected to an external stress, such as acid or osmotic shocks48, 52, 53. Furthermore,for some species acidification of the cytoplasm has been shown to be related to pathogenicity54, 55, and in yeast changes in thecytoplasmic pH affect particle cytoplasm26. Here we investigate if surface attachment methods influence the internal pH ofbacteria during time lapse imaging.

3/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

To monitor the effect of the adhesive substances on the internal pH of E. coli cells during their growth on the coatedsurfaces, we use a genetically encoded indicator pHluorin16, 56, 57. pHluorin is a variant of the green fluorescent protein withpH sensitive spectrum that responds in a ratiometric manner, SI Figure 4. Prior to the growth experiments, pHluorin hasbeen carefully calibrated both in vivo and in vitro, see Supplementary Materials16. For the in vivo calibration we’ve usedvarious ∆pH = pHexternal − pHinternal collapsing agents and noticed that the calibration curves deviate slightly depending onthe uncoupler, which compromises the accuracy of the potential pH measurements. Though it is not clear what causes thedifference in the calibration curves, we show that the combination of potassium benzoate and methylamine hydrochloride(PBMH) allows us to reproduce the in vitro calibration most accurately (SI Figure 5), and we subsequently use PBMH for invivo calibration.

Having calibrated pHluorin, we measure the intracellular pH of the immobilised bacteria during growth and divisiontracking two generations, as shown in Figure 3. We find the average intracellular pH equals 8.39± 0.33 on all the testedsurfaces at the beginning of experiments (this is the average pH value of the first 70 min in Figure 3A). For cells grown inMM9 the cytoplasmic pH decreases to 7.79 in an exponential-like manner over the course of several hours, thus we obtain thesteady-state pH value by fitting all the pH traces in MM9 with exponential decay function. The intracellular pH drops mainly inthe first generation and then remains costant with some cell to cell variation shown in SI Figure 6. In contrast, cytoplasmic pHof the cells growing in LB starts from 8.36±0.06 and slightly increases within the two generations, SI Figure 7.

The intracellular pH values we measured are higher than those commonly found in the literature (7.2-7.847, 48). We assumethese discrepancies originate from the fact that we constantly exchange the medium during the cell growth, removing fromthe environment metabolic waste products and any quorum sensing or signal molecules, which have been shown to influencecytoplasmic pH before49, 58, 59. Indeed, when cells are kept in the original growth medium their cytoplasmic pH varies between7.3 and 7.7, Figure 3B. It, however, increases rapidly to 8.2-8.3 when fresh medium is supplied and can be reduced back to∼7.8 upon the return to the original growth medium, Figure 3B. Further incubation in a fresh medium with no exchange (flowhas been stopped) leads to the pH decrease to ∼ 7.5 after ∼ 10 min.

A B

Figure 3. Intracellular pH dynamics of bacterial cells immobilised on a given surface. (A) Single cell intracellular pH ismeasured with cytoplasmic pHluorin during cell growth on different surfaces. Solid line and shaded area show the mean andstandard deviation. All tested immobilisation methods exhibit similar tendencies: pH values start from pH 8.38±0.29 anddecrease to about pH 7.78±0.16 within ∼7 h of observation. The inset shows all single-cell traces plotted for each condition.(B) Cells in a flow chamber are sequentially treated with the medium taken from a growing culture (blue regions) and freshmedium (white regions). Media exchange occurs in a short pulse manner at the beginning of each period, while no flow isapplied during imaging. Data points show the mean of N > 500 measured bacteria with standard deviation as error bar.

Attachment quality on different surfaces variesFor single cell imaging it is important that analysed cells remain ”flat”22 (long axis parallel to the imaging plane) for theduration of observation, which could last several generations. Of the surfaces we have tested, ”PLL in-chamber”, ”PLL rinsed”,”PLL air-dried”, PEI and Cell-Tak show similar attachment quality during cell growth and division. These surfaces also showconsistent imaging quality under phase-contrast microscopy, Figure 4. Cells on the gel pad surface also remain ’flat’, but weobserve agarose impurities that influence image quality, Figure 4. Typical time lapse images can be done on these surfaces for

4/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

PLLin-chamber

PLLrinsed

PLLair-dried

PEI

Cell-Tak

Gel pad

Figure 4. Examples of phase contrast images of E. coli cells attached to the surface with different immobilisation methods(scale bar is 5 µm). All images shows good ”flatness” (see Materials and Methods for definition) and contrast for cellmorphology tracking.

few generations. Covalent binding to APTES coated surface resulted in attached E. coli cells, however the cells usually detachduring the cell growth leading to unsuccessful cell morphology tracking and pH measurements.

DiscussionGood surface attachment is an important requirement for bacterial single-cell studies using optical microscopy. However, weare unaware of a systematic study that characterises the effects on cells’ physiology caused by different adhesives. Changes inthe cellular physiology caused by different surface attachment methods can influence not only studies of cellular physiologythemselves, but also studies focusing on specific cellular molecular mechanisms. For example, metabolic rate or internal pHcould lead to the alteration of cytoplasm properties, e.g. its fluidity25, 26, and many intracellular processes, including DNAreplication and cell division, are highly dependent on the growth rate60, 61. It is, therefore, important to consider and characterisepotential effects of the immobilisation method on physiology of the studied bacteria.

Here, we test a range of the immobilisation techniques and show that E. coli’s growth rate and shape are immobilisationmethod independent. Cell length and the growth rate are dependent on the growth medium, as expected, but independent of thesurface attachment chemistry. Length ratio during the cell cycle (L f /L0) stays constant for all tested conditions, as has beenpreviously noticed62, 63. However, we see evidence of adaptation to the attachment to the surface itself, which is a relevantfinding given the importance of understanding the physiology of surface-attached bacteria64.

Interestingly, we notice that the cell size of bacteria growing on the surface is reduced compared to the free swimmingbacteria, though it doesn’t seem to be correlated with the attachment method. We speculate mechanosensing is involved to anextent, which could be further tested by using strains lacking pili and flagella (such as YD13317) that are known to be involvedin surface sensing in bacteria65.

Though the concerns regarding use of PLL for surface attachment have been previously expressed in the literature29, 66,we note that the experiments that demonstrate inhibition of cell division by PLL, do so for the case of free PLL molecules inthe medium30. We show that all three of the tested PLL-coating protocols leave no residual PLL in the medium and do notinfluence bacterial growth rate and division.

For measurements of E. coli’s cytoplasmic pH we use pHluorin and find that in vivo calibration curve is dependant onthe agents used to collapse pH. We do not understand the observations at present, but speculate that it could occur due to theinteraction between the uncoupler (CCCP or indole) and the bacterium, e.g. CCCP can be actively exported by EmrAB-TolCpump67. Using pHluorin we show that the internal pH of the attached E. coli is kept between 7.3 and 8.4 and doesn’t varysignificantly with the surface coating. On all the tested surfaces in MM9 media the cytoplasmic pH decreases slightly (from

5/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

pH 8.38 to 7.78) in the course of the experiment, and the behaviour changes in LB medium, where cytoplasmic pH startsat pH 8.29 and increases over the observation time to pH 8.43. The result is not unexpected as metabolic byproducts havebeen demonstrated to influence E. coli’s cytoplasmic pH49, 58, 68–71. For example, glucose metabolism mainly produces organicacids as byproducts, such as acetate, lactate, formate, succinate etc.68. These organic acids are capable of crossing the innermembrane in their uncharged form dissociating in the cytoplasm, which causes full or partial collapse of the pH gradient acrossthe membrane49, 58, 69. In the case of LB medium, the alkalinisation of the media due to E. coli’s metabolism has been reportedand attributed to the release of the amine-containing compounds70, 71.

We, thus, conclude that all the tested bacteria immobilisation protocols can be used for live cells imaging without affectingcells’ main physiological traits.

Materials and Methods

Bacterial Strains and Growth ConditionsThe strain, EK03, is the Escherichia coli K-12 MG1655 strain with ”sticky” flagella mutation16, 72 and pkk223-3/pHluorin(M153R)plasmid. The plasmid containing pHluorin with M153R mutation to have better stability of fusion proteins56 was a kind giftfrom Dr. Tohru Minamino.

Cell cultures were inoculated in MM9 medium (Na2HPO4 50 mM, NaH2PO4 20 mM, NaCl 8.5 mM, NH4Cl 20 mM, CaCl20.1 mM, KCl 1 mM, MgSO4 2 mM, 0.5% glucose , and MEM Amino acides solution (Gibco, USA) with 1:100 dilution froman overnight culture grown in Lysogeny broth (LB: 1% w/v Bacto-Tryptone, 0.5% w/v yeast extract and 0.5% w/v NaCl).Cultures are grown for 4 h at 37C with continuous shaking (220 rpm) to reach optical density values between 0.4 and 0.5. Thecells were further diluted 1:60 into MM9 or LB for the surface immobilisation protocols.

Microscope and Microfluidic chambersBacterial growth assays were conducted using a motorised, inverted optical microscope (Ti-E, Nikon, Japan) with perfectfocus system for time lapse observation. The microscope is equipped with a 100x Objective (Plan Apo 100x/1.45NA lambda,Nikon, Japan), sCMOS Camera (Zyla 4.2, Andor, UK) and LED fluorescent excitation light source (PE4000, CoolLED, UK).Imaging was performed in phase-contrast and epifluorescence configuration, the latter was used for measuring the cytoplasmicpH with pHluorin. The exposure times for phase-contrast and epifluorescence imaging were 100 and 70 ms respectively, andimages were recorded every 5 min. The excitation wavelengths imaging of pHluorin were 395 and 470 nm, achieved via adual-band dichroic mirror (403/502nm, FF403/502-Di01-25x36, Semrock, USA) and a dual-band bandpass emission filter(FF01-433/530-25, Semrock, USA).

To supply fresh oxygenated media throughout the experiment we use a flow chamber made as follows. Two 1.5 mm holeswere drilled in a microscope slide 20 mm apart. PTFE tubing with inner diameter 0.96 mm was attached to the slide with epoxyglue, SI Figure 8. The flow chamber was then created by attaching double sided tape or gene frame (Fisher Scientific, USA) tothe slide and covering it with pre-coated or uncoated cover glass depending on the immobilisation protocol. Gene frame wasused to create a larger chamber to fit the agarose pad, while sticky tape was used for all of the other protocols. Dimensionsof the formed flow chamber are 3.5×25×0.2 mm for doubled sided tape, and 17×28×0.25 mm for gene frame. The flowchamber construction protocol varied slightly with different coating protocols. For the PLL ”in chamber” and Cell-Tak coatingmethods, the flow chambers were sealed before coating. In other cases, the coverslips were coated prior to the flow chamberconstruction.

For all of the immobilisation assays medium was flown at 400 µl/min flow rate at the end of the attachment protocol for4 min to remove poorly attached cells, upon which the the flow rate was altered to 4 µl/min for the duration of the experiment(12 h). Media was flown with a syringe pump (Fusion 4000, Chemyx, USA ).

Immobilisation protocolsPreparation step: coverslip cleaningThe coverslip is sonicated in an ultrasonic bath with saturated solution of KOH in ethanol for 30 min. It is then rinsed with thedeionised water and sonicated again. Glass treated this way does not allow cells attachment unless it’s coated. The cleaningstep has been performed prior to all attachment protocols.

PLL ”in-chamber”Surface of the flow chamber is coated with 0.1% poly-L-lysine (PLL) by flushing PLL through the channel for ∼10 s followedby washing it out with the excessive volume of growth medium (x20 times the volume of the chamber). Cells are then loadedinto the flow chamber and incubated for 1 to 3 min to allow attachment, and then washed out as described under Microscopyand Microfluidic chambers section.

6/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

PLL ”rinsed”Coverslip is coated with the PLL prior to the flow chamber construction. 100 µl of the 0.01% PLL solution (diluted from P8920Sigma-Aldrich, USA ) is spread over approximate 1.5 cm2 area. The solution is allowed to sit on the coverslip for 30 min, thenwashed off with 5 ml of deionised water. The coated coverslip is then used to construct a flow chamber. The cells are attachedas above.

PLL ”air-dried”The protocol is similar to the ”rinsed” method. Here, the PLL solution on the coverslip is air dried fully, typically for 1.5 h. Thecoated coverslip is then washed with 5 ml deionised water and used to construct a flow chamber. The cells are attached asabove.

PEI200 µl of 1% PEI is spread out on the coverslip covering the area the size of the flow chamber tunnel. Solution is incubated onthe surface for 10 s, and washed off thoroughly with 100 ml of deionised water. The volume of the water should be much higherthan that of the PEI to leave no residual PEI molecules that are not attached to the glass surface. Otherwise (e.g. if we use 5 mlof water), we observe cells blebbing73, 74 and dying when grown on the surface. We also notice that the ”in-chamber” coatingmethod is not applicable for PEI, as it leads to MM9 precipitation in the chamber caused by the leftover free PEI molecules.The cells are subsequently attached as described above.

Cell-TakCell-Tak working mixture is prepared by adding 14 µl of Cell-Tak (1.16 mg/mL) to 174 µl NaHCO2 (pH 8.0) followed byimmediate vortexing. A pre-assembled flow chamber is incubated with Cell-Tak mixture for 20 min, then washed with 3 ml ofdeionised water. 3 ml of MM9 is flushed through the chamber before attaching the cells. Finally, cells are attached as above.

APTESA coverslip was incubated in 2% APTES for 2 h, and then rinsed it with deionised water and acetone. The remaining acetonewas air-dried with nitrogen. The coated coverslip was later used for the flow chamber construction. Cells are attached as above.

Agarose Gel PadFor the agarose gel pad, the flow chamber area was 17×28 mm2. The gel pad was created by adding a 5 µl droplet of melted1% agarose to the middle of the flow chamber. 0.5 µl of the cell culture was added onto the solidified pad and covered with thecoverslip. The fresh medium was constantly circulated around the agarose ”island” during the experiment.

Image analysisCell segmentationPhase contrast images of the cells were analysed with custom written Python script and ImageJ75. In phase microscopy, cellsappear as dark objects on a light background, with a characteristic white halo, Figure 1 inset. Cells are segmented withWatershed algorithm76, 77 implemented as ImageJ Marker-controlled Watershed plugin78. The algorithm treats the image asa geological terrain. The lighter watershed which separates neighboring darker drainage basins is the edge around region ofinterest. After the segmentation, the cell length is calculated by PSICIC algorithm (Projected System of Internal Coordinatesfrom Interpolated Contours)79. Briefly, the algorithm finds two poles of a cell as points that are the greatest Euclidean distanceapart, thus creating two curves. On each of the two contour curves the algorithm evenly distributes equal number of points andthen connects them (effectively creating width lines). Finally, the center line, i.e. the lenght of the cell, runs along the middle ofthe width lines79.

Growth curve fittingThe growth of the cell follows L = L0 × ebt + c, where L0 is the cell length at the start of the recording, t time in minutes,b the growth rate, and c the constant representing the length of non-growing poles. The equation is fitted to the cell lengthusing Levenberg-Marquardt algorithm80, a non-linear least squares method, implemented using Scipy, a numerical package inPython81. Initial parameters we identified by first fitting a polynomial to the logarithmic growth of the cell length.

References1. Dobell, C. Antony van Leeuwenhoek and his ”Little animals” (Harcourt, Brace and company, 1932).

2. Elowitz, M. B., Levine, A. J., Siggia, E. D. & Swain, P. S. Stochastic Gene Expression in a Single Cell. Sci. (80-. ). 297,1183–1186 (2002).

7/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

3. Taheri-Araghi, S., Brown, S. D., Sauls, J. T., McIntosh, D. B. & Jun, S. Single-Cell Physiology. Annu. Rev. Biophys. 44,123–142 (2015). DOI 10.1146/annurev-biophys-060414-034236. arXiv:1011.1669v3.

4. Raskin, D. M. & de Boer, P. A. J. Rapid pole-to-pole oscillation of a protein required for directing division to the middle ofEscherichia coli. Proc. Natl. Acad. Sci. 96, 4971–4976 (1999). DOI 10.1073/pnas.96.9.4971.

5. Ghodke, H., Ho, H. & van Oijen, A. M. Single-molecule live-cell imaging of bacterial DNA repair and damage tolerance.Biochem. Soc. Trans. 46, 23–35 (2018). DOI 10.1042/BST20170055.

6. Hol, F. J. H. & Dekker, C. Zooming in to see the bigger picture: Microfluidic and nanofabrication tools to study bacteria.Sci. (80-. ). 346 (2014).

7. Leygeber, M. et al. Analyzing microbial population heterogeneity - expanding the toolbox of microfluidic single-cell cultivations. J. Mol. Biol. (2019). URL http://www.sciencedirect.com/science/article/pii/S0022283619302323. DOI https://doi.org/10.1016/j.jmb.2019.04.025.

8. Wang, P. et al. Robust Growth of Escherichia coli. Curr. Biol. 20, 1099–1103 (2010).

9. Humphries, J. et al. Species-Independent Attraction to Biofilms through Electrical Signaling. Cell 168, 200–209.e12(2017). DOI 10.1016/j.cell.2016.12.014.

10. Okumus, B. et al. Single-cell microscopy of suspension cultures using a microfluidics-assisted cell screening platform.Nat. Protoc. 13, 170–194 (2018). DOI 10.1038/nprot.2017.127.

11. Hadizadeh Yazdi, N., Guet, C. C., Johnson, R. C. & Marko, J. F. Variation of the folding and dy-namics of the Escherichia coli chromosome with growth conditions. Mol. Microbiol. 86, 1318–1333(2012). URL http://www.ncbi.nlm.nih.gov/pubmed/23078205http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC3524407. DOI 10.1111/mmi.12071.

12. Moffitt, J. R., Lee, J. B. & Cluzel, P. The single-cell chemostat: An agarose-based, microfluidic de-vice for high-throughput, single-cell studies of bacteria and bacterial communities. Lab Chip 12, 1487–1494(2012). URL http://www.ncbi.nlm.nih.gov/pubmed/22395180http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC3646658. DOI 10.1039/c2lc00009a.

13. Priest, D. G., Tanaka, N., Tanaka, Y. & Taniguchi, Y. Micro-patterned agarose gel devices for single-cell high-throughputmicroscopy of E. coli cells. Sci. Rep. 7, 17750 (2017). DOI 10.1038/s41598-017-17544-2.

14. Lo, C.-J., Sowa, Y., Pilizota, T. & Berry, R. M. Mechanism and kinetics of a sodium-driven bacterial flagellar motor. Proc.Natl. Acad. Sci. U. S. A. 110, E2544–E2551 (2013). DOI 10.1073/pnas.1301664110.

15. Rosko, J., Martinez, V., Poon, W. & Pilizota, T. Osmotaxis in Escherichia coli through changes in motor speed. Proc. Natl.Acad. Sci. U. S. A. 114, E7969–E7976 (2017). DOI 10.1073/pnas.1620945114. 1703.03926.

16. Krasnopeeva, E. Single cell measurements of bacterial physiology traits during exposure to an external stress. Ph.D. thesis,The University of Edinburgh (2018).

17. Pilizota, T. & Shaevitz, J. W. Fast, multiphase volume adaptation to hyperosmotic shock by Escherichia coli. PLoS One 7,e35205 (2012). DOI 10.1371/journal.pone.0035205.

18. Meyer, R. L. et al. Immobilisation of living bacteria for AFM imaging under physiological conditions. Ultramicroscopy(2010). DOI 10.1016/j.ultramic.2010.06.010.

19. Sydor, A. M., Czymmek, K. J., Puchner, E. M. & Mennella, V. Super-Resolution Microscopy: From Single Molecules toSupramolecular Assemblies. Trends Cell Biol. 25, 730–748 (2015). DOI 10.1016/j.tcb.2015.10.004.

20. Coltharp, C. & Xiao, J. Superresolution microscopy for microbiology. Cell. Microbiol. 14, 1808–1818 (2012).

21. Pilizota, T. & Shaevitz, J. Plasmolysis and cell shape depend on solute outer-membrane permeability during hyperosmoticshock in e. coli. Biophys. J. 104, 2733 – 2742 (2013). URL http://www.sciencedirect.com/science/article/pii/S0006349513005614. DOI https://doi.org/10.1016/j.bpj.2013.05.011.

22. Buda, R. et al. Dynamics of escherichia coli’s passive response to a sudden decrease in external osmolarity. Proc.Natl. Acad. Sci. 113, E5838–E5846 (2016). URL https://www.pnas.org/content/113/40/E5838. DOI10.1073/pnas.1522185113. https://www.pnas.org/content/113/40/E5838.full.pdf.

23. Tipping, M. J., Steel, B. C., Delalez, N. J., Berry, R. M. & Armitage, J. P. Quantificationof flagellar motor stator dynamics through in vivo proton-motive force control. Mol. Microbiol. 87,338–347 (2013). URL https://groups.physics.ox.ac.uk/molecularmotors/pdfsofpapers/MolMicroTippingpRPMFstators.pdf. DOI 10.1111/mmi.12098.

8/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

24. Kapanidis, A. N., Uphoff, S. & Stracy, M. Understanding protein mobility in bacteria by tracking single molecules.J. Mol. Biol. 430, 4443 – 4455 (2018). URL http://www.sciencedirect.com/science/article/pii/S0022283618303905. DOI https://doi.org/10.1016/j.jmb.2018.05.002. Plasticity of Multi-Protein Complexes.

25. Parry, B. R. et al. The bacterial cytoplasm has glass-like properties and is fluidized by metabolic activity. Cell 156,183–194 (2014). DOI 10.1016/j.cell.2013.11.028. NIHMS150003.

26. Munder, M. C. et al. A pH-driven transition of the cytoplasm from a fluid- to a solid-like state promotes entry intodormancy. Elife 5 (2016). DOI 10.7554/eLife.09347.

27. Cowan, S. E., Liepmann, D. & Keasling, J. D. Development of engineered biofilms on poly- l-lysine patterned sur-faces. Biotechnol. Lett. 23, 1235–1241 (2001). URL https://doi.org/10.1023/A:1010581503842. DOI10.1023/A:1010581503842.

28. Touhami, A., Jericho, M. H., Boyd, J. M. & Beveridge, T. J. Nanoscale characterization and determination ofadhesion forces of Pseudomonas aeruginosa pili by using atomic force microscopy. J. Bacteriol. 188, 370–377(2006). URL http://www.ncbi.nlm.nih.gov/pubmed/16385026http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC1347306. DOI 10.1128/JB.188.2.370-377.2006.

29. Katsu, T., Tsuchiya, T. & Fujita, Y. Dissipation of membrane potential of Escherichia coli cells induced by macromolecularpolylysine. Biochem. Biophys. Res. Commun. 122, 401–6 (1984).

30. Colville, K., Tompkins, N., Rutenberg, A. D. & Jericho, M. H. Effects of poly(L-lysine) substrates on attached Escherichiacoli bacteria. Langmuir 26, 2639–2644 (2010). DOI 10.1021/la902826n.

31. Conte, M., Aliberti, F., Fucci, L. & Piscopo, M. Antimicrobial activity of various cationic molecules on foodbornepathogens. World J. Microbiol. Biotechnol. 23, 1679–1683 (2007). URL http://link.springer.com/10.1007/s11274-007-9415-6. DOI 10.1007/s11274-007-9415-6.

32. Bakshi, S., Siryaporn, A., Goulian, M. & Weisshaar, J. C. Superresolution imaging of ribosomes and RNA polymerase inlive Escherichia coli cells. Mol. Microbiol. 85, 21–38 (2012).

33. Mohapatra, S., Choi, H., Ge, X., Sanyal, S. & Weisshaar, J. C. Spatial Distribution and Ribosome-Binding Dynamics ofEF-P in Live Escherichia coli. MBio 8, e00300–17 (2017).

34. Wang, S., Arellano-Santoyo, H., Combs, P. a. & Shaevitz, J. W. Measuring the bending stiffness of bacterial cells using anoptical trap. J. Vis. Exp. 8–9 (2010). DOI 10.3791/2012.

35. Dickson, J. S. & Koohmaraie, M. Cell surface charge characteristics and their relationship to bacterial attachment to meatsurfaces. Appl. Environ. Microbiol. 55, 832–836 (1989). URL http://aem.asm.org/.

36. Sowa, Y. et al. Direct observation of steps in rotation of the bacterial flagellar motor. Nat. 437, 916–919 (2005). URLhttp://www.nature.com/articles/nature04003. DOI 10.1038/nature04003.

37. Pilizota, T. et al. A molecular brake, not a clutch, stops the rhodobacter sphaeroides flagellar motor. Proc.Natl. Acad. Sci. 106, 11582–11587 (2009). URL https://www.pnas.org/content/106/28/11582. DOI10.1073/pnas.0813164106. https://www.pnas.org/content/106/28/11582.full.pdf.

38. Sowa, Y. & Berry, R. M. Bacterial flagellar motor. Q. Rev. Biophys. 41, 103–132 (2008). URL https://groups.physics.ox.ac.uk/molecularmotors/pdfsofpapers/QRBSowaBFMreview2008.pdf.DOI 10.1017/S0033583508004691.

39. Waite, J. H. & Tanzer, M. L. Polyphenolic substance of Mytilus edulis: Novel adhesive containing L-dopa and hydroxypro-line. Sci. (80-. ). 212, 1038–1040 (1981). DOI 10.1126/science.212.4498.1038.

40. Papov, V. V., Diamond, T. V., Biemann, K. & Waite, J. H. Hydroxyarginine-containing polyphenolic proteins in the adhesiveplaques of the marine mussel Mytilus edulis. J. Biol. Chem. 270, 20183–20192 (1995). DOI 10.1074/jbc.270.34.20183.arXiv:1011.1669v3.

41. Lee, B. P., Messersmith, P., Israelachvili, J. & Waite, J. Mussel-inspired adhesives and coatings. Annu. Rev. Mater.Res. 41, 99–132 (2011). URL https://doi.org/10.1146/annurev-matsci-062910-100429. DOI10.1146/annurev-matsci-062910-100429. https://doi.org/10.1146/annurev-matsci-062910-100429.

42. Gokaltun, A., Yarmush, M. L., Asatekin, A. & Usta, O. B. Recent advances in nonbiofouling PDMSsurface modification strategies applicable to microfluidic technology. TECHNOLOGY 05, 1–12 (2017).URL http://www.ncbi.nlm.nih.gov/pubmed/28695160http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC5501164. DOI 10.1142/s2339547817300013.

9/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

43. Harris, L. K. & Theriot, J. A. Surface Area to Volume Ratio: A Natural Variable for Bacterial Morphogenesis. TrendsMicrobiol. 26, 815–832 (2018). URL http://www.ncbi.nlm.nih.gov/pubmed/29843923http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC6150810. DOI 10.1016/j.tim.2018.04.008.

44. Gorden, J. & Small, P. L. Acid resistance in enteric bacteria. Infect. Immun. 61, 364–7 (1993).

45. Lin, J., Lee, I. S., Frey, J., Slonczewski, J. L. & Foster, J. W. Comparative analysis of extreme acid survival in Salmonellatyphimurium, Shigella flexneri, and Escherichia coli. J. Bacteriol. 177, 4097–4104 (1995). DOI 10.1128/jb.177.14.4097-4104.1995.

46. Slonczewski, J. L., Fujisawa, M., Dopson, M. & Krulwich, T. A. Cytoplasmic pH Measurement and Homeostasis inBacteria and Archaea. Adv. Microb. Physiol. 55 (2009). DOI 10.1016/S0065-2911(09)05501-5.

47. Zilberstein, D., Agmon, V., Schuldiner, S. & Padan, E. Escherichia coli intracellular pH, membrane potential, and cellgrowth. J. Bacteriol. 158, 246–252 (1984).

48. Slonczewski, J. L., Rosen, B. P., Alger, J. R. & Macnab, R. M. pH homeostasis in Escherichia coli: measurement by 31Pnuclear magnetic resonance of methylphosphonate and phosphate. Proc. Natl. Acad. Sci. U. S. A. 78, 6271–6275 (1981).DOI 10.1073/pnas.78.10.6271.

49. Lund, P., Tramonti, A. & De Biase, D. Coping with low pH: Molecular strategies in neutralophilic bacteria. FEMSMicrobiol. Rev. 38, 1091–1125 (2014). DOI 10.1111/1574-6976.12076.

50. Mitchell, P. Coupling of phosphorylation to electron and hydrogen transfer by a chemi-osmotic type of mechanism. Nat.191, 144–148 (1961). DOI 10.1038/191144a0.

51. Bearne, S. L. Illustrating the effect of pH on enzyme activity using Gibbs energy profiles. J. Chem. Educ. 91, 84–90 (2014).DOI 10.1021/ed400229g.

52. Martinez, K. A. et al. Cytoplasmic pH response to acid stress in individual cells of Escherichia coli and Bacillussubtilis observed by fluorescence ratio imaging microscopy. Appl. Environ. Microbiol. 78, 3706–3714 (2012). DOI10.1128/AEM.00354-12.

53. Chakraborty, S., Winardhi, R. S., Morgan, L. K., Yan, J. & Kenney, L. J. Non-canonical activation of OmpR drives acidand osmotic stress responses in single bacterial cells. Nat. Commun. 8, 1587 (2017). DOI 10.1038/s41467-017-02030-0.

54. Miller, S. I., Kukral, A. M. & Mekalanos, J. J. A two-component regulatory system (phoP phoQ) controls Salmonellatyphimurium virulence. Proc. Natl. Acad. Sci. U. S. A. 86, 5054–5058 (1989). DOI 10.1073/pnas.86.13.5054.

55. Choi, J. & Groisman, E. A. Acidic pH sensing in the bacterial cytoplasm is required for Salmonella virulence. Mol.Microbiol. 101, 1024–1038 (2016). DOI 10.1111/mmi.13439.

56. Morimoto, Y. V., Kojima, S., Namba, K. & Minamino, T. M153R mutation in a pH-sensitive green fluorescent proteinstabilizes its fusion proteins. PLoS One 6, e19598 (2011). DOI 10.1371/journal.pone.0019598.

57. Krasnopeeva, E., Lo, C.-J. & Pilizota, T. Single-cell bacterial electrophysiology reveals mechanisms of stress-induced damage. Biophys. J. 0 (2019). URL https://linkinghub.elsevier.com/retrieve/pii/S0006349519303923http://arxiv.org/abs/1809.05306. DOI 10.1016/j.bpj.2019.04.039. 1809.05306.

58. Wilks, J. C. et al. Acid and base stress and transcriptomic responses in bacillus subtilis. Appl. Environ. Microbiol. 75,981–990 (2009). URL https://aem.asm.org/content/75/4/981. DOI 10.1128/AEM.01652-08. https://aem.asm.org/content/75/4/981.full.pdf.

59. Zarkan, A. et al. Indole Pulse Signalling Regulates the Cytoplasmic pH of E. coli in a Memory-Like Manner. Sci. Rep. 9,3868 (2019). URL http://www.nature.com/articles/s41598-019-40560-3. DOI 10.1038/s41598-019-40560-3.

60. Wallden, M., Fange, D., Lundius, E. G., Ozden Baltekin & Elf, J. The synchronization of replication and divisioncycles in individual e. coli cells. Cell 166, 729 – 739 (2016). URL http://www.sciencedirect.com/science/article/pii/S0092867416308601. DOI https://doi.org/10.1016/j.cell.2016.06.052.

61. Thomas, P., Terradot, G., Danos, V. & Weiße, A. Y. Sources, propagation and consequences of stochasticity in cellulargrowth. Nat. Commun. 9, 4528 (2018). URL http://www.nature.com/articles/s41467-018-06912-9.DOI 10.1038/s41467-018-06912-9.

62. Osella, M., Nugent, E. & Cosentino Lagomarsino, M. Concerted control of Escherichia coli cell division. Proc. Natl. Acad.Sci. U. S. A. 111, 3431–5 (2014). DOI 10.1073/pnas.1313715111.

10/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

63. Campos, M. et al. A constant size extension drives bacterial cell size homeostasis. Cell 159, 1433–1446 (2014). DOI10.1016/j.cell.2014.11.022. NIHMS150003.

64. Beloin, C., Roux, A. & Ghigo, J. M. Escherichia coli biofilms. Curr. Top. Microbiol. Immunol.322, 249–89 (2008). URL http://www.ncbi.nlm.nih.gov/pubmed/18453280http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC2864707http://www.ncbi.nlm.nih.gov/pubmed/18453280%0Ahttp://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC2864707.

65. Ellison, C. K. et al. Obstruction of pilus retraction stimulates bacterial surface sensing. Sci. 358, 535–538 (2017). URLhttps://science.sciencemag.org/content/358/6362/535. DOI 10.1126/science.aan5706. https://science.sciencemag.org/content/358/6362/535.full.pdf.

66. Cattoni, D. I., Fiche, J. B., Valeri, A., Mignot, T. & Nollmann, M. Super-Resolution Imaging of Bacteria in a MicrofluidicsDevice. PLoS ONE 8 (2013). DOI 10.1371/journal.pone.0076268.

67. Griffith, J. M. et al. Experimental evolution of escherichia coli k-12 in the presence of proton motive force (pmf) uncouplercarbonyl cyanide m-chlorophenylhydrazone selects for mutations affecting pmf-driven drug efflux pumps. Appl. Environ.Microbiol. 85 (2019). URL https://aem.asm.org/content/85/5/e02792-18. DOI 10.1128/AEM.02792-18.https://aem.asm.org/content/85/5/e02792-18.full.pdf.

68. Bennett, B. D. et al. Absolute metabolite concentrations and implied enzyme active site occupancy in Escherichia coli.Nat. Chem. Biol. 5, 593–599 (2009). DOI 10.1038/nchembio.186.Absolute.

69. Kitko, R. D. et al. Cytoplasmic acidification and the benzoate transcriptome in bacillus subtilis. PLOS ONE 4, 1–12 (2009).URL https://doi.org/10.1371/journal.pone.0008255. DOI 10.1371/journal.pone.0008255.

70. Lazar, S. W., Almiron, M., Tormo, A. & Kolter, R. Role of the Escherichia coli SurA protein in stationary-phase survival.J. Bacteriol. 180, 5704–5711 (1998). DOI D - NLM: PMC107631 EDAT- 1998/10/29 MHDA- 1998/10/29 00:01 CRDT-1998/10/29 00:00 PST - ppublish.

71. Molina, P. M., Parma, A. E. & Sanz, M. E. Survival in acidic and alcoholic medium of Shiga toxin-producing Escherichiacoli O157:H7 and non-O157:H7 isolated in Argentina. BMC Microbiol. 3, 1–6 (2003). DOI 10.1186/1471-2180-3-17.

72. Kuwajima, G. Construction of a minimum-size functional flagellin of Escherichia coli. J. Bacteriol. 170, 3305–3309(1988). DOI 10.1128/jb.170.7.3305-3309.1988.

73. Yao, Z., Kahne, D. & Kishony, R. Distinct Single-Cell Morphological Dynamics under Beta-LactamAntibiotics. Mol. Cell 48, 705–712 (2012). URL http://www.ncbi.nlm.nih.gov/pubmed/23103254http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC3525771. DOI10.1016/j.molcel.2012.09.016.

74. Fletcher, E., Pilizota, T., Davies, P. R., McVey, A. & French, C. E. Characterization of the effects of n-butanol on the cellenvelope of E. coli. Appl. Microbiol. Biotechnol. 100, 9653–9659 (2016). URL http://link.springer.com/10.1007/s00253-016-7771-6. DOI 10.1007/s00253-016-7771-6.

75. Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. methods 9, 676 (2012).

76. Meyer, F. & Beucher, S. Morphological segmentation. J. Vis. Commun. Image Represent. 1, 21–46 (1990). URLhttps://doi.org/10.1016/1047-3203(90)90014-m. DOI 10.1016/1047-3203(90)90014-m.

77. Soille, P. Morphological image analysis: principles and applications (Springer Science & Business Media, 1999).

78. Legland, D., Arganda-Carreras, I. & Andrey, P. MorphoLibJ: integrated library and plugins for mathematical morphologywith ImageJ. Bioinforma. btw413 (2016). URL https://doi.org/10.1093/bioinformatics/btw413. DOI10.1093/bioinformatics/btw413.

79. Guberman, J. M., Fay, A., Dworkin, J., Wingreen, N. S. & Gitai, Z. PSICIC: Noise and asymmetry in bacterial divisionrevealed by computational image analysis at sub-pixel resolution. PLoS Comput. Biol. 4 (2008). DOI 10.1371/jour-nal.pcbi.1000233.

80. More, J. The levenberg-marquardt algorithm: Implementation and theory. In Watson, G. (ed.) Numerical Analysis, vol.630 of Lecture Notes in Mathematics, 105–116 (Springer Berlin Heidelberg, 1978). URL http://dx.doi.org/10.1007/BFb0067700. DOI 10.1007/BFb0067700.

81. Jones, E., Oliphant, T., Peterson, P. et al. SciPy: Open source scientific tools for Python (2001–). URL http://www.scipy.org/.

11/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

82. Chimerel, C., Field, C. M., Pinero-Fernandez, S., Keyser, U. F. & Summers, D. K. Indole prevents Escherichia colicell division by modulating membrane potential. Biochim. Biophys. Acta - Biomembr. 1818, 1590–1594 (2012). DOI10.1016/j.bbamem.2012.02.022.

AcknowledgementsThis project is supported by the Human Frontier Science Program Grant (RGP0041/2015) to TP, CJL and FB. CJL is financiallysupported by the Ministry of Science and Technology, Republic of China under contract No. MOST-107-2112-M-008-025-MY3.EK was supported by the Global Research and Principal’s Career Development PhD Scholarships. We thank Tom Shimizu andVictor Caldas (AMOLF) for sharing the APTES protocol of immobilization.

Author contributions statementYKW, EK, FB, TP and CJL designed research. YKW and EK performed research and analysed data. YKW, EK, FB, TP andCJL interpreted results and wrote the paper. All authors reviewed the manuscript.

12/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

Supplementary MaterialsSupplementary Methods: pHluorin calibrationThe in vivo calibration of pH sensor was performed as follows16. The mixture of 100 mM MES, HEPES and AMPSO bufferswas adjusted to a set of pH values in the range between 5.5 and 9, and supplemented with one of the three pH collapsingagents: 40 mM potassium benzoate and 40 mM methylamine hydrochloride (PBMH)52, 25 µM CCCP or 5 mM indole82.Tunnel-slides were prepared as previously described16, 57: two bits of sticky tape form a tunnel and are sandwiched betweena coverslip and a microscope slide. Buffer of known pH was flushed into a channel, incubated for 15 min, upon which5 different fields of view containing over 100 cells were imaged with 50 ms exposure time. The calibration curves wereplotted as ratio of emission intensities for excitation at 395 nm and 475 nm against pH, and fitted with the sigmoid functionR395/475 = (a1ek(pH−pH0)+a2)/(ek(pH−pH0)+1), where a1, a2, k and pH0 are free fitting parameters.

In vitro calibration was performed with the purified pHluorin protein diluted into buffer of known pH in the 96-well plate(Thermo Scientific, Optical bottom). The pHluorin excitation spectra for 510 nm emission was measured in Spark 10Mmultimode plate reader (Tecan Trading AG, Switzerland). The His-tagged protein was purified using affinity chromatographycolumn16. The excitation spectra was scanned from 380 nm to 480 nm with 5 nm step size. Additionally, the autofluorescenceof the buffer with no added protein was measured and subtracted from the measured protein intensity.

Supplementary Figures

PLLin-chamber

PLLrinsed

PLLair-dried

PEI Cell-Tak Gel pad0.0

0.5

1.0

1.5

2.0

2.5

3.0

Lf/L

0

SI Figure 1. Final length to the initial length ratio (L f /L0) stays constant when cells are attached to the surface with differentimmobilization methods. L0 is the length after the first division, and L f the length before the second division.

13/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

0.000

0.005

0.010

0.015

0.020

0.025

0.030Gr

owth

rate

(1/m

in)

0

2

4

6

8

Initi

al le

ngth

(m

)

Gel padMM9

Gel padLB

0

2

4

6

8

Gel padMM9

Gel padLB

0

1

2

Wid

th (

m)

A

C

BLf/L

0

D

SI Figure 2. Comparison of (A) the growth rate, (B) initial cell length L0, (C) length ratio L f /L0 and (D) cell width for cellsgrowing on the gel pad in MM9 and LB media.

0 50 100 150 200 250 300 350 400 450 500Time(min)

1

2

3

4

5

6

7

Leng

th d

istrib

utio

n(m

)

SI Figure 3. Cell length distribution over time on the PLL air-dried surface. The average length of the cells population changesfrom 3.53±0.92 µm to 2.53±0.74 µm. A steady length distribution is reached by the second generation of surface-grownbacteria with the mean being roughly 28% smaller than that of the liquid cultured bacteria.

14/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

380 400 420 440 460 480Wavelength (nm)

0

20

40

60

80

100

Inte

nsity

(a.u

.)

pHluorin 5.5 pHluorin 7 pHluorin 9

SI Figure 4. Excitation spectra of purified pHluorin at pH values 5.5 (red), 7.0 (yellow) and 9.0 (green). Emission is collectedat 510 nm.

5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0pH

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

395/

475

nm ra

tio

In vitroPBMHIndoleCCCP

SI Figure 5. Comparison of the in vivo and in vitro calibration curves. Coloured lines show in vivo pHluorin calibrationcurves with 40 mM PBMH (red), 5 mM indole (yellow) or 25 µM CCCP (green) as ∆pH collapsing agents. Black curve showsin vitro calibration curve of purified pHluorin in buffer with no supplements. In vivo curve with PMBH aligns best with the invitro curve.

15/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

0 100 200 300 400 500 600Time(min)

4

5

6

7

8

9

10

11

12

13

pH

Generation 1Generation 2

SI Figure 6. Single cell intracellular pH dynamics for two generations of bacteria, first generation is shown in green and thesecond in red. Cells are grown on the PLL ”in-chamber” coated surface.

0 100 200 300 400 500 600Time(min)

6

7

8

9

10

11

pH

LB Gel padMM9 Gel pad

SI Figure 7. Mean and the standard deviation of the intracellular pH of the cells grown in the gel pad in LB (green) or MM9(red) media.

16/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint

Agaroseisland

A B

C

SI Figure 8. Schematic of our flow chambers for bacterial immobilsation experiments. (A) top view of the flow chamber usedfor surface attachments, (B) top view of chamber for agarose island experiments, and (C) side view of flow chambers.

17/17

.CC-BY-NC-ND 4.0 International licenseacertified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under

The copyright holder for this preprint (which was notthis version posted May 24, 2019. ; https://doi.org/10.1101/648840doi: bioRxiv preprint